Traumatic brain injury (TBI) is a disorder of public health concern that has ramifications not only for the individual, but also for their family, the public health system, and the economy. The majority of TBIs affect people in the prime of their vocational productivity. Despite decades of work in the area of cognitive rehabilitation, a recent evidence report summarizing research on the efficacy of cognitive remediation after TBI revealed predominately negative results. This is partly because there is currently no empirically supported theory about cognitive recovery from TBI to guide intervention strategies. This study will provide better scientific evidence to guide cognitive remediation by more fully characterizing the potential early learning mechanisms of TBI patients, and by prospectively evaluating the recovery trajectories of both automatic and controlled cognitive processes. TBI patients will complete experimental tasks designed to assess automatic and more controlled components of visual search, semantic priming, and memory. These task will be administered following the TBI patients emergence from post-traumatic amnesia, and then again at 2-, 6-, and 12-month intervals. Control participants will complete the tasks at similar intervals following baseline testing. Perceptually-based implicit learning abilities and memory-based skill learning abilities will also be evaluated acutely following injury. If restitution of function of automatic processes occurs early in recovery and before controlled processes, then training techniques that tap into residual automatic skills or capitalize on processes that can be made automatic through practice could prove vital for early interventions, facilitating recovery and improving ultimate cognitive outcome. Furthermore, an understanding of the recovery trajectories of automatic processes could help refine models for predicting rehabilitative gains and aid in rehabilitative planning and resource allocation as automatic processes often serve as a "data base" for more controlled processes.